通过选择性跳过不确定样本对齐，提升开集半监督学习的泛化与新类检测能力。
Let the Void Be Void: Robust Open-Set Semi-Supervised Learning via Selective Non-Alignment
- 对低置信度未标注样本采用选择性跳过对齐，仅保留与已知类原型的排斥力。
- 在多个数据集上实现比现有方法更高的未知类检测率，且不降低已知类分类精度。
- 适合需要鲁棒泛化和新类别发现的开放环境应用，如自动驾驶、医疗诊断。
开集半监督学习（OSSL）利用包含已知分布（ID）和未知分布（OOD）样本的未标注数据，同时提升闭集分类准确率并检测新出现的未知类。现有方法或丢弃不确定样本的信息，或强制将所有未标注样本对齐至少数几个合成的“万能”表示，导致几何坍缩并在已知未知类上产生过度自信。为解决此问题，本文提出选择性非对齐机制，在对比学习的传统拉近（pull）与推远（push）操作中引入新的“跳过”算子。所提出的框架SkipAlign，选择性跳过低置信度未标注样本的对齐操作，仅保留其与已知类原型的温和排斥。这一策略使不确定样本转化为纯粹的排斥信号，从而形成更紧凑的已知类簇，同时让未知类特征自然分散。大量实验表明，SkipAlign在不牺牲已知类分类性能的前提下，显著优于当前最优方法，在未见未知类检测上表现更优。
原文摘要 · Abstract (English)
Open-set semi-supervised learning (OSSL) leverages unlabeled data containing both in-distribution (ID) and unknown out-of-distribution (OOD) samples, aiming simultaneously to improve closed-set accuracy and detect novel OOD instances. Existing methods either discard valuable information from uncertain samples or force-align every unlabeled sample into one or a few synthetic "catch-all" representations, resulting in geometric collapse and overconfidence on only seen OODs. To address the limitations, we introduce selective non-alignment, adding a novel "skip" operator into conventional pull and push operations of contrastive learning. Our framework, SkipAlign, selectively skips alignment (pulling) for low-confidence unlabeled samples, retaining only gentle repulsion against ID prototypes. This approach transforms uncertain samples into a pure repulsion signal, resulting in tighter ID clusters and naturally dispersed OOD features. Extensive experiments demonstrate that SkipAlign significantly outperforms state-of-the-art methods in detecting unseen OOD data without sacrificing ID classification accuracy.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。